Related Experiment Video
Updated: Sep 30, 2025
08:31
A Simplified Stepwise Approach to Echo Guidance during Percutaneous Mitral Valve Repair
Published on: October 16, 2021
4.0K
A system for real-time multivariate feature combination of endoscopic mitral valve simulator training data
Reinhard Fuchs1, Karel M Van Praet2,3, Richard Bieck4
1Innovation Center Computer Assisted Surgery, University of Leipzig, Leipzig, Germany. reinhard.fuchs@medizin.uni-leipzig.de.
Summary
This study developed a multimodal system to automatically assess surgical skills during endoscopic training. The system effectively differentiates between 2D and 3D views using motion and muscle data, achieving over 90% accuracy.
Area of Science:
- Medical Simulation
- Surgical Skill Assessment
- Human-Computer Interaction
Background:
- Traditional endoscopic training evaluation is time-consuming.
- Automated surgical skill assessment is needed for efficient training.
- Current methods lack simulator independence and fixed metrics.
Purpose of the Study:
- Analyze learning benefits of stereoscopic (3D) view in endoscopic training.
- Develop a custom, simulator-independent surgical evaluation system.
- Establish a multimodal system for objective skill assessment.
Main Methods:
- Performed data fusion of motion and muscle-action measurements.
- Collected data from experts with varying skill levels using 2D and 3D imaging.
- Calculated training features and assessed significance using distance and variance analysis.
- Utilized features for automatic classification of endoscope modes.
Main Results:
- Recorded 324 datasets from 12 volunteers, including spatial and electromyographic (EMG) data.
- Identified amplitude-related muscle information and hand/wrist velocity as significant features.
- Achieved >90% accuracy in classifying the endoscope view mode (2D vs. 3D).
Conclusions:
- The setup and feature calculation are valid for distinguishing endoscopic view modes.
- Significant distinctions in features aid in identifying the endoscopic view.
- This work is a step towards real-time, automated endoscopic training evaluation and progress tracking.
